I think one or more of the contributions to this thread expressed dismay that by referencing an I-D from a registry (or approving a registry policy that permits the same) we harm ourselves by giving the lie to the "inappropriate to use Internet-Drafts as reference" I-D boilerplate. Others have pointed out the "including informal documentation” phrasing of RFC 8126’s definition of “Specification Required”. These two things are manifestly in tension, but I’d like to point out that to resolve that tension by being sticklers for the draft boilerplate disclaimer, i.e. refusing to approve registries that permit I-D’s as references for Specification Required, would be to create a perverse incentive for authors to do their work outside the IETF. After all, we have no similar disclaimer associated with, say, a post on a person’s LiveJournal account, so by keeping their “informal documentation” out of the I-D space, they would give themselves greater freedom of action. I don’t see how doing this would benefit the IETF community. AFAICT the people working on 8126bis are well aware of this issue and we should wait and see what they propose as a resolution, in the meantime muddling along with the imperfect world we have. —John
